Decline of Science: Computer Science and Databases

        I've read with interest your three last columns, and thought you might want to know an exemplary field to analyse the decline of knowledge.

        I work with databases. I'm not particularly good, in part because I haven't a formal CS education, but I try to educate myself.

        And what I learned scares me.

        You may know that more than 30 years ago an IBM Research Fellow — as in basic research — redefined the database field. The name is Eduard F "Ted" Codd, and he did it by taking a totally ad hoc field to the new highs of a solid foundation on set theory and first order predicate logic. The rest is History.

        Or it should be.

        Actually, you may also know that Mr Codd's relational model for database management was seldom properly implemented, and never widely deployed. Moreover, the bastard version that did get publicised has damaged practitioners even more and worse than COBOL and BASIC did for programmers, and the situation there was already bad enough that your recent columns remind me a lot of Edsger Wybe Dijkstra's writings.

        One would think that after blundering for 30 years and seeing the successful frustrations or outright failures of SQL, OODBs, and whatever, the field would naturally look for the relational flame as ignited by Codd and now kept by Christopher J Date & followers. Actually, being proved wrong scares people so much, including founders of companies that should be in touch with the technology, that no one wants to hear. As publication and seminars is mostly now done only if sponsors can be found, the relational explainers are shut out of mindshare.

        It gets even uglier. Mr Date, being famous, can live passably. Others either have independent incomes or find themselves with little income, thus having to either leave the field or compromise their intellectual integrity.

        To topple it all with a further connection to your recent columns, there is now a company patenting some basic technologies on the implementation of relational databases. They are very secretive, so things that should have been published at least four or five years ago are only now coming to light, and only as patents applications which aren't known for clearness. And if their business model fail, as it seems it will, these patents could end up locked in the virtual vault of some SQL vendor and left there until the Earth and skies are redone or the patent expires, whatever comes first.

        In fairness it must be said that a small Utah company recently has finally implemented and made generally available a faithful implementation of the relational model. Not being free software nor portable, at least not yet — they yet don't "get" free software but are striving for portability —, the impact they will have in markets and minds remains to be seen, as well — again — their business model viability.

        So I think here we have a field which is fundamental to IT, and therefore to modern society, and not only is languishing but actually regressing.

        Should you think it interesting, I could provide you some pointers.

-- 
  _


/ \ Leandro Guimarães Faria Corsetti Dutra        +41 (21) 216 15 93


\ / http://homepage.mac.com./leandrod/        fax +41 (21) 216 19 04
  X  http://tutoriald.sourceforge.net./      Orange Communications CH


/ \ Campanha fita ASCII, contra correio HTML      +41 (21) 644 23 01